Discussion: MapAnt - Switzerland and Liechtenstein

https://www.mapant.ch/

mamma mia!

mamma mia - da vero !

Nice. And after wondering what the large "forbidden area" in the East is: it's the Swiss National Park, the only area of level "IUCN Ia Strict Nature Reserve" in Switzerland.

Little comparison, same patch processed using mapant.no algoritm, actual O map and mapant.ch. Apparently they have decided to make it a lot more generalized way compared to the other mapant projects. I think generalized result is visually more pleasing and more uniform, but usability for regular orienteering style activities is quite limited. Still most likely in a coutry with terrains like that it may be good idea to process everything with generalized/fast way and as a step 2 select areas for more detailed processing. Just to avoid wasting too much resources for analysing areas unsuitable for orienteering. Maybe they have make this next step at some point.

I assume the main difference is that they're using the OCAD algorithm for these tiles.

Partly maybe, but I think it is also deliberate, because there is 5m contours with no formlines and no attempts to map cliffs. I am sure skilled 0cad used would bake something more delicate in no-time. In case somebody feels like testing, this is the file
https://data.geo.admin.ch/ch.swisstopo.swisssurfac...

Does the CH algorithm really get all those trails that the NO algorithm misses? Or are the trails taken from some other source?

Trails and such are not detected from lidar, those are simply drawn from other sources. I did not bother donload/draw any additional vectors.

And there would be an extensive database of those in Switzerland, particularly the (many) designated walking routes. (I was impressed that it picked up the single track along the Rhone gorge downstream of Geneva that's an occasional running destination of mine when I'm there).

Trails are identical here, both draw them from the same database I guess.
https://map.geo.admin.ch

I ran the Lidar from Jagge's link through my map generation script to see how it compares. Using my default settings, my map comes out with less contour detail than the mapant.no version but more detail than the mapant.ch version. And, there are differences where my algorithm inserts dots for small hills. But, the big difference is in the vegetation classification, which my scripts generalize into something that looks much more like the real orienteering map.

Here's a link to the map samples: https://www.dropbox.com/sh/q80soiuyl57phyv/AAB3Ilu...

Some wags like to claim that in the Australian winter we have a greater area of snow cover than Switzerland. I think Swiss wags might be able to claim that their country has more contours than Australia.

That claim is quite plausible - really it's just a reflection of the fact that Switzerland is a very small country

It depends on how you do your counting. Australia has 448 contours. Switzerland has 888.

It also depends on what you define as "area of snow cover". I suspect it wouldn't hold up if you defined it as, say, the area which was under snow for at least 1 month each winter. Switzerland's area is 41,000 square kilometres of which I'd guess 50-70% would meet that criterion. Australia has 6,000 square kilometres above 1500m, which is probably a reasonable approximation for the winter snowline under that definition (at least in the main mountainous areas).

(It may have been more true in the past when the snowline was lower - you take in a lot more area if you drop to 1300 or 1400m).

I figured it meant that at some moment in the Australian winter, there's more snow coverage there than there is in Switzerland (which is experiencing summer at that point).

No it means there is a surprisingly large area of snow coverage in Australia w hether or not is actually larger than Switz .

Most winters there would be at least one day when the area covered by snow in Australia is larger than the total area of Switzerland, but it rarely lasts long outside the highest elevations - on the mainland, it would be unusual for snow to stay on the ground for more than 24 hours at any elevation below 1000m.

1500 m is pretty high for an arbitrary cut off elevation. It has been nearly 20 years since I was last there (don't know what it is like these days) but Lake Mountain had reliable snow and operated as a CC ski resort. Its highest point is below 1500m and the trail head is around 1350m I think. There was occasionally no snow right at the trailhead but always snow higher up. Certainly it would have been very rare indeed for there to be no snow (even at the trailhead) for more than a month during the 'ski season'.

But that assertion about Australia and Switz has been around for a long time - so it would be based on assumptions of lower elevations of snow cover. I seem to recall reading somewhere that Harrietvlle (around 500m asl) got regular snow back in the 19th century.

Anyway whether it is literally true is beside the point - the point of the comparison is to show that there is actually a very large area that gets regular snow in Australia during the winter - something non Australians (and even many Australians) are surprised to learn.

The phrases "Lake Mountain" and "reliable snow cover" are not seen in conjunction all that often these days. Same for Dingo Dell/Cresta. The latter hasn't been repaired since fires. Wondering if Selwyn will suffer a similar fate.

Well - as I said - that Switzerland comparison goes back a bit - remember hearing it in the 80's. When I was doing a fair it of skiing (90's) I always thought of the "snowline" as around 1300m.

BTW I was trying to find Victorian topos online to look at Lake Mountain - couldn't find anything - just a site offering sale of pdfs. Is there a site where you can actually see them - VicMaps?

Oddly enough, the best place I've found for online Victorian topography is the Victorian Emergency website - if you zoom in enough on the map you get topographic layers.

Thanks Blair. So how many km2 in Australia over 1300m?

And where do you find that out?

Geoscience Australia, but they only do it in 500-metre bands.

Thanks again.

So the reason you choose 1500m was not because it was the best fit but because it was the only statistic you had available?

I hope you don't take that approach when writing climate change papers ;)

My go to for Victorian on line topos is MapshareVic. You can get 10 metre contours from that (or 5 in metro Melbourne). I think Vic Emergency only does 20 metres. Still frustrating compared to the MapAnt products.
https://mapshare.vic.gov.au/MapShareVic/

Ian's map looks really great! But I wouldn't blame Ocad's granulated vegetation too much, it sort of gives impression how distictive edge there may be. With polished result it is difficult to guess. But I guess that is needed only if veg interpretation is not "good enough".

Thanks Neil

1500 was a guesstimate (before I started looking for areas). At a national level I suspect it's fairly close to the mark - offsetting slightly lower levels in Victoria against higher ones in areas further north.

The vegetation mapping in both the '.ch' and '.no' outputs seem to have the issue that they don't show direct changes from 'open' to 'dense'. It looks like both algorithms always transition from 'open' to 'white' to 'light green' to darker greens. I think that comes from doing averaging across the boundary, but it doesn't seem like it represents a runner's experience in many cases.

What it comes to no, green and yellow are entirely different algorithms and green and yellow can overlap (yellow overwrites green). Requirements for yellow most likely makes that white edge, maybe all yellow parches should be a tad larger for leaf on data, but that might make it not map single trees of mountain areas with white blobs with leafs off data, so it ended up like that.